Please. Help. Trying to chose a software, not sure how to begin...

#1
H,

I was originally going to do a confirmatory factor analysis (CFA) on some data, but it was recommended to me that I conduct an exploratory structural equation model (ESES) instead by a very knowledgeable faculty member.

I was directed towards using MPLUS, however, it is very expensive.

Can anybody direct me towards any free software that could perform an ESES? Or perhaps let me know whether the free MPLUS demo version would be sufficient?

I don't know much about ESES at this point, including where or how to begin the process if I can't even select a software program to start working with...

Thank you very much for your time, and even more so for potentially helpful responses.

All the best.
 
#3
The first question I would even start with is actually asking yourself if you really are doing ESEM vs. an EFA model. The two are often confused and in the end I see most users are really just doing EFA, not ESEM. If all you are modelling is a measurement model, then ESEM=EFA for the most part (with minor differences in rotations etc.,). You won't need an expensive software program like MPlus to do that. I would start there first. What are you actually modelling, that will determine if a EFA or ESEM is appropriate.
 
#4
The first question I would even start with is actually asking yourself if you really are doing ESEM vs. an EFA model. The two are often confused and in the end I see most users are really just doing EFA, not ESEM. If all you are modelling is a measurement model, then ESEM=EFA for the most part (with minor differences in rotations etc.,). You won't need an expensive software program like MPlus to do that. I would start there first. What are you actually modelling, that will determine if a EFA or ESEM is appropriate.
Thanks for your reply!

With regards to your question, I was planning on doing an ESES model on a 24-item (5-level Likert type) questionnaire.

The questionnaire I'll be examining was constructed from 8-items taken from three already existing scales; wherein they did a forced three-factor principal components analysis with orthogonal varimax rotation and the eight items with highest factor loadings on Factor I were determined, then Factors II and III were labeled. And a single high-loaded factor was disclosed as well.

Anyways, I was originally going to do a CFA since this publication already existed, but I was encouraged to do an ESES for reasons I don't fully understand just yet, something about how as a field we're not as ready for CFAs as we often think we are...
 
#5
Thanks for your reply!

With regards to your question, I was planning on doing an ESES model on a 24-item (5-level Likert type) questionnaire.

The questionnaire I'll be examining was constructed from 8-items taken from three already existing scales; wherein they did a forced three-factor principal components analysis with orthogonal varimax rotation and the eight items with highest factor loadings on Factor I were determined, then Factors II and III were labeled. And a single high-loaded factor was disclosed as well.

Anyways, I was originally going to do a CFA since this publication already existed, but I was encouraged to do an ESES for reasons I don't fully understand just yet, something about how as a field we're not as ready for CFAs as we often think we are...
I'd run an EFA (can do this in SPSS), then a CFA (Mplus is my preference, but many software can cope with a CFA) afterwards.